TR&D 4. Algorithms for functional and anatomical brain analysis Principal investigators: Susumu Mori, Professor of Radiology Michael Miller, Professor of Biomedical Engineering SUMMARY The overall objective of TR&D4 is to develop cutting-edge image analysis technologies that can integrate various anatomical representations (multi-manifold) of the brain based on multi-modal MRI that includes not only multiple MRI contrasts (e.g. T1, T2, diffusion, susceptibility imaging, CEST imaging) but also functional MRI (fMRI), physiological MRI (measures of cerebral blood volume, CBV, flow, CBF and oxygen metabolism, CMRO2), and MR spectroscopy (MRS) into a common anatomical framework. As TR&D 1-3 further develop new image acquisition and signal processing techniques, the richness of the modern MRI-based observations continues to grow with more contrasts, higher spatial dimensions, as well as the addition of a time axis, which eventually needs to be correlated with clinical features. For example, stratification of patients by anatomical, functional, and spectroscopic scans could play a crucial role in many collaborative projects for brain development, ADHD, schizophrenia, and Huntington's Disease. For neonatal, Alzheimer's disease, multiple sclerosis, and primary progressive aphasia studies, anatomical and diffusion scan results need to be integrated over the longitudinal dimension. For certain applications, there are specific structures of interest such as the limbic system, for which detailed surface shape analysis based on high-resolution atlases are needed. The various image analysis tools we have developed in the past 15 years are now being used by investigators around the world. These tools are now widely distributed through www.mristudio.org with more than 9,000 registered users, the number of processed data sets in our server exceeded 10,000 in 2014 alone, a number that is still increasing. For the next five years, our specific aims are: Aim 1: Integration of diffeomorphic brain mapping and multi-atlas based image feature analysis We will integrate cutting-edge multi-atlas fusion algorithms into our LDDMM-based pipelines and develop a new image segmentation platform through the integration of tools developed in Aim 2. We will also continue to develop brain atlas resources through continued development of multi-atlas libraries that covers the entire age range. Aim 2: Integrative analysis for multi-contrast and multi-manifold mapping of the whole brain anatomy Tools to integrate multiple image contrasts and multiple shape objects will be developed for advanced analysis, which will include the modern MRI data developed in TR&D 1-3, high-resolution atlases based on 7T/11.7T scanners, and tools for time dependent anatomical change analysis. Aim 3: Deployment and Maintenance of Cloud-based service system Toward the sunset of this National Resource Grant, it is important to establish a sustainable platform for dissemination and services. In 2015, we introduced a new platform-free web-based cloud system as our new service approach, which drastically reduces the burden for users and developers. We bridge publicly available supercomputing resources and users who need the computational power. This platform will also be used to deploy software programs developed in TR&D 1?3 including MRS, rs-fMRI, QSM, CBV, CBF, and CMRO2 analyses.